We propose an intervention to teach preschoolers with autism to attend to, interpret and respond to visual communication, in order to improve their opportunities to profit, as typically developing children do, from the rich social learning opportunities offered in the natural environment. The intervention will use the child's gaze as input, and training structured according to the principles of applied behavior analysis, delivered in a voluntary play context, without imposing any restraint on the child's freedom to leave the environment if and when s/he chooses. This training is directed at attracting and increasing the child's gaze at faces, and laying the groundwork for development of functional face gaze, by associating positive consequences to the child's gaze at the faces of Virtual Buddies. The child's automatically monitored direction of gaze will result in contingent changes in the display, additional opportunities to acquire the target behavior and, on reaching criterion, a reward that may include social praise and display of enthusiasm from a Virtual Buddy, and the Buddy's invitation to enjoy the ensuing visual display, auditory display, and/or a ride, since the 'pod' in which the child sits is an adapted kiddie-ride helicopter. The development and pilot testing proposed here will yield information necessary to the design of clinical trials. Objectives include design and implementation of a virtual environment (VE), to include a virtual world with both animated and static objects including vehicles, people, animals and toys, a cast of Virtual Buddies, including one representing each child's sibling, with a rich repertoire of behaviors and facial expressions, and a training protocol to reinforce gaze at social stimuli and provide experience of positive consequences to attending to and successfully interpreting communication transmitted by facial expression. We will test the system in young typically developing children and children with autism, and then conduct a preliminary study to define appropriate values of independent variables for a Pilot Study. The Pilot Study will use single case methodology to assess learning, retention and generalization within the VE, and carry-over to the natural environment.